Guide on Gender Analysis of Census Data Full Draft of 6 December 2012 Contents
Country example 13: Brazilian Census Using Decision-maker to Designate the Household Head
Download 4.06 Mb.
|
- Navigate this page:
- Methodology Box 6: Mixing Different Levels of Analysis
- 4. Tabulations
- Table 22: Timor Leste - Population in private households by marital status according to sex and relationship to the head of household
- Table 23: Cambodia 2008 - Average numbers of household members classified by age and sex of the head of household and by relationship to the head of household (all numbers multiplied by 100)
Country example 13: Brazilian Census Using Decision-maker to Designate the Household Head The 2010 Brazilian census included the possibility of more than one household head (“one person”/“more than one person”) which could be considered as shared responsibilities in the household – but without using the term ‘household head’, as this term has been considered increasingly inappropriate. In the early 90’s, surveys conducted by the NSO (IBGE) started using the term “reference person”; later, after pilot tests, it was replaced with the term “responsible person”, which was widely accepted by respondents. The 2000 and 2010 censuses incorporated this change but didn’t establish a criterion to the selection of household “responsible person”. In the 2010 census of Brazil, the “responsible person” is chosen among the household residents. Preliminary results of 2010 Census show that 70.4% of the households have only one “responsible person”. The questionnaire of the 2011 census of South Africa provided a definition of the household head as “the person who is the main decision-maker in the household”. It considered the possibility of joint responsibility, in which case it instructed census interviewers to take the oldest person. 284. Despite these limitations, the head of household measure has been used frequently in gender studies to document the erosion of male exclusive providers, and has given new meaning to this aspect of domestic living. An example of this ambivalence is illustrated in the World’s Women 2010 (United Nations, 2010 a), which cautions that generalizations between “female-headed households” and “male-headed households” are impossible because of the contextual differences in women’s and men’s statuses as well as the variety of headship definitions and the households that may be included under these labels. Therefore, researchers must proceed cautiously to compare across households headed by women and men, in order to understand patterned poverty and vulnerability in different types of household structures. This caution often translates into additional precision in the analysis and household specification, as the World’s Women 2010 then reports that Latin American households headed by lone mothers with children have higher poverty rates than those of lone fathers with children, and that poverty rates for women living in one-person households are higher than the corresponding rates for men.
285. Because of the problems involved, the household head concept should be clarified, and perhaps it should be replaced by several concepts to provide a better understanding of household structure and processes. Statisticians and scholars alike have made the case to clarify the concept of head of household. Although most censuses seek to define one concept of household head, it would actually help the analysis if the different concepts were treated separately. As Budlender (1997: 15) states: Instead of insisting on one definition, one could several questions so as to get at multiple definitions of household and multiple axes of household headship, and then leave it to users/analysts of the data to pick up the particular definitions they find useful in addressing a specific question. Multiple questions would also allow for cross-tabulation and comparison of the definitions, to see where they differ, why they differ and by how much. These questions are not only of academic interest. They are also of interest to policy-makers who are eager to target their programmes where they will be most effective, and at those most in need, while avoiding the allocation of scarce resources to less ‘appropriate’ beneficiaries’. Failure to clarify the concept of head of households has a number of potentially undesirable consequences for the presentation and interpretation of results – specifically in the areas of 1) what is being measured; 2) problems for analysis and cross-country comparisons; and 3) gender stereotypes. 286. Some countries, like Ireland and Aruba, have replaced the simple relationship with the household head by a more complex matrix of relationships between household members, but this can be time-consuming to fill out, even though the matrix is limited only to relationships with the first four household members, as follows: Relationship of person 5 (or 6, 7, etc.) to persons 1 2 3 4 1. Husband or wife _ _ _ _ 2. Partner (including same-sex partner) _ _ _ _ 3. Son or daughter _ _ _ _ 4. Step-child _ _ _ _ 5. Brother or sister _ _ _ _ 6. Mother or father _ _ _ _ 7. Grandparent _ _ _ _ 8. Step-mother or father _ _ _ _ 9. Son or daughter-in-law _ _ _ _ 10. Grandchild _ _ _ _ 11. Other related _ _ _ _ 12. Unrelated (including foster child) _ _ _ _ 287. Aruba came up with an alternative approach to classify family nuclei and household types in its 2010 population census. Through a set of questions it was possible to identify the categories and sub-categories of household type and family type. Moreover, the questions allowed making a household classification using both a formal approach (married couples) as well as a sociological approach (married + consensual unions). The questions used to make this classification were:
This approach worked well, especially with digital questionnaires, where the person number of related people could be chosen from an answer box containing the names of all eligible candidates. However, Aruba is a small country and it is doubtful whether the same technique could be used on a larger scale in censuses in much bigger countries. 288. The lack of detail, and inconsistency between countries among the relationship with head categories in the census presents another limitation. International recommendations support the collection of detailed information on relationships, with a sufficient number of pre-coded categories. Nevertheless, in some countries the number of options is extremely limited (e.g. only 5 relationship categories in Malawi and Bangladesh). As was mentioned in Chapter 3, some censuses ask about the father and mother of each person residing in the household, i.e. a) If the father/mother of the person reside in the same household (e.g. Aruba, Barbados, Cape Verde, South Africa); and b) If so, the identification of the person's father/mother. This information can be quite helpful to analyse residential patterns, e.g. by providing an indicator of the likely distribution of unpaid care work, because in many countries, more children live with their mothers than with their fathers. The 2006 census of the Maldives went even further and asked for each child under the age of 18 whether he/she lived with:
2. Mother and stepfather; 3. Father and stepmother; 4. Mother only; 5. Father only; 6. Other relatives; 7. Only unrelated household members.
2. Spouse; 3. Stepchildren; 4. Other relatives; 5. Unrelated individuals; 6. Without guardian. 4. Tabulations 289. The Principles and Recommendations consider household and family characteristics as an essential topic to be investigated and suggest that NSOs should construct the following tabulations relevant for gender analysis: Recommended tabulations for population censuses:
Additional tabulations for population censuses:
290. To illustrate (part of) the first kind of table, one may take the 2004 census of Timor Leste as an example, as displayed in Table 22. Among other things, this table shows that a surprisingly high percentage of female heads of households (38.1 per cent) are actually married; only the number of widowed female heads of household (48.5 per cent) is larger. Most of the remaining female heads of household are single, with very few cases of divorce or separation. Note also that couples are about as likely to live with the parents of the woman (1,385) as with the parents of the man (1,410). Surprisingly, in almost half of the former (602) cases, the couple is not formally married and in 113 cases the man continues to live with his parents-in-law even after the wife has died. The latter construction is even more common in the case of widowed daughters-in-law (379). Mothers live with their children much more often (7,008 cases) than fathers (1,994). Table 22: Timor Leste - Population in private households by marital status according to sex and relationship to the head of household
Source: Population and Housing Census of Timor Leste (2004): Table 3.1 291. The following is an illustration of another kind of table that can be extracted from a census (in this case, the 2008 census of Cambodia) by combining information about the sex and age of the head of household with information on the relationship of the household members with the head of household. Table 23: Cambodia 2008 - Average numbers of household members classified by age and sex of the head of household and by relationship to the head of household (all numbers multiplied by 100)
292. Some of the conclusions that can be drawn from this table – at least for the case of Cambodia – are the following:
293. As was mentioned before, much use has been made of the criterion of whether the head of household is male or female. Some of the limitations of this distinction were discussed earlier in this sub-chapter. Tabulating census data on the basis of male/female headship only gives a first impression of the social and economic position of women, but should be complemented by the comparison of the structure of households. Sometimes the percentage of women in the household is used as a criterion. For example, The World’s Women 2010 (United Nations, 2010 a: 159) mentions that “households with an overrepresentation of women might be more likely to be found below the poverty line”, but immediately points out that this may be due to two rather different causes, namely: 1) in some types of households where the share of women is higher, the earnings per capita tend to be lower because women’s participation in the labour market and their earnings are lower than men’s; and 2) as the ratio of women to men increases with age, the presence of non-earning older persons in extended households may depress per capita household income. While the first explanation points to a genuine gender disparity, the second is a demographic composition factor that one would like to control for, rather than confounding it with the first. 294. When using household information for gender analysis, special attention should be dedicated to the situation of elderly women compared to elderly men. The situation of elderly household members varies considerably among countries. In Brazil, for example, the presence of an elderly person with a retirement pension or allowance was found to reduce the likelihood of family vulnerability to poverty, and having an elderly person in the family was even strategically more effective in reducing vulnerability than having a spouse (Lavinas and Nicoll, 2007). The structure of gender imbalances at older ages may be different from the typical pattern at younger ages and care must be taken not to assume automatically that all of them will favour men over women. As Knodel and Ofstedal (2003: 693) point out, In many settings, perhaps even in most, older women may be disadvantaged relative to older men on some or most dimensions of well-being. Clearly, too, there are numerous exceptions as indicated by the examples presented above. A major hindrance to making a more definitive statement about gender and aging is that systematic empirical assessments comparing the situation of older men and women are inadequate for drawing a firm conclusion-especially for the developing world, where the large majority of older persons live. Most broad statements asserting a generalized female disadvantage in old age appear to be based on presumptions and an incomplete allowance for the full set of influences over the life course, including later stages, that determine well-being in old age. In any event, generalizations about which sex is more disadvantaged are of limited value. To more fully understand the effects of gender on the well-being and support of older persons and their families, research must move beyond assumptions of universal gender inequality and the disadvantaged situation of older women to examine the experiences of older men and women within the contexts in which they live. Such research should recognize that well-being at older ages is multidimensional and that gender differences may go in either direction or, for that matter, be largely absent, depending on the aspect of well-being under consideration. 295. Migration is another significant factor that has to be controlled, for any comparisons to be meaningful. Klasen, Lechtenfeld and Povel (2010), found, for example, that in Thailand and Viet Nam households with female headship that had a male family member residing elsewhere were, on average, better off than the general population, whereas female-headed households without such external remittances were poorer and more vulnerable than average. --------------------------------------------------------- Download 4.06 Mb. Share with your friends:
|